Voice activity detectors (VAD) are used in microphone applications to monitor input and determine when intended speech is or is not occurring. The VAD determination of voice or no voice may be used in digital signal processing (DSP) voice processing algorithms which adapt filters to noise for transmit signal (Tx) noise reduction. The VAD allows the voice processing algorithms to adapt the noise filters only when speech is not present.
In the prior art, typical VADs detect speech by analyzing the input signal received at the microphone. For example, the signal level of the input signal may be measured and compared to a pre-determined threshold level above which speech is determined to be occurring and below which speech is determined not to be occurring.
Voice activity detectors known in the prior art may also detect speech using an external sensor (also referred to herein as a VAD sensor) such as an accelerometer in contact with a wearer's head. The VAD sensor, using appropriate software and hardware, indicates when speech is occurring based on detection of tissue vibration associated with human speech by the wearer. However, one problem with the prior art VAD sensors is that they must be in complete contact with the user head in order to function. If complete contact is not present, the VAD sensor does not function properly. As a result, any application relying on the VAD sensor determination does not function properly. For example, the aforementioned DSP noise filtering algorithm does not perform as desired when the voice activity detection determination is inaccurate.
Prior art VAD sensors typically use some form of a mechanical means to ensure that the sensor is in contact with the user skin. However, neither the user nor any subsequent processing algorithm is provided any feedback whether the VAD sensor is properly positioned. In a noise reduction application, the Tx noise reduction will not function if the user that does not position the VAD sensor correctly. In some cases, improper positioning of the VAD may prevent the Tx operation from functioning completely.
As a result, there is a need for improved methods and apparatuses for improved voice activity detection.